Transitioning transmitters to COFDM

With the continued debate over 8VSB and COFDM, making a transmitter purchase decision has become even more difficult. Perhaps it is time to take a minute to ask, "What if?" Can stations that have already invested in an 8VSB transmitter simply swap out an exciter and go about their business, or is there another shoe to drop? It's important to know the longer-reaching effects of any possible change, especially one that comes at a make-or-break point for the entire industry.

Toward that end, we posed several questions to two industry leaders, Geoff Mendenhall, vice president of advanced product development for Harris Broadcast, and Don Markley, BE's own Transmission & Distribution columnist and president of Markley and Associates. The questons were as follows:

"Suppose the FCC did allow COFDM, what effect would that have on transmitters already purchased? Can modern DTV transmitters be converted to COFDM operation? Will there be a power penalty? Are there other long-term issues to consider if a station needs to make the transmitter decision now? Should I wait or buy now?"

In light of the continued debate over 8VSB and COFDM, should I wait to order a transmitter?

No. If you wait until the last minute, you may jeopardize your ability to meet the FCC's mandated on-air date; making logistics much more difficult for your station, and end up paying more than you would by moving ahead on a realistic implementation schedule. To this end, we would recommend going ahead and getting on the air now with 8VSB. In the unlikely event the transmission standard changes, you can convert your transmitter to COFDM. We don't believe there will be a standards change in the U.S., but if you are concerned, you may want to size key RF components for the peak power required by COFDM as a precaution. This would support conversion to COFDM if necessary, with minimum risk, while preserving the major portion of the initial investment. The cost of replacing the DTV exciter will be small compared to the overall cost of the transmitting facility, and Harris transmitter architectures are scaleable.

- Suppose the FCC did allow COFDM, what effect would that have on transmitters already purchased?

If COFDM were permitted in the U.S., VHF and UHF transmitters already installed and operating with 8VSB could be converted by changing the 8VSB exciter to COFDM while maintaining the current peak RF power output within the ratings of the current transmitter. It may also be necessary to adjust some metering, control and power amplifier parameters of the transmitter.

- Could today's 8VSB transmitters be converted to COFDM operation?

Yes, by replacing the existing 8VSB exciter with a 6MHz COFDM exciter and making other miscellaneous adjustments including the linearization of the RF power amplifiers for the different peak-to-average ratio of the COFDM signal. Harris is a leading supplier of COFDM transmission equipment to the European DVB-T market and has significant expertise in COFDM. All Harris solid-state and IOT DTV transmitters have been laboratory tested with COFDM modulation and are compatible with both 8VSB and COFDM modulation.

- Would there be an RF power penalty in switching from 8VSB to COFDM?

Yes. Assuming that it is important to replicate existing analog coverage, there will be a significant power penalty. COFDM has higher RF power peaks than 8VSB. Harris has confirmed that even after the application of aggressive crest factor reduction techniques, the peak-to-average power ratio is at least 2dB greater for COFDM. This means that a transmitter running at full power with 8VSB will deliver 2dB less average power after conversion to COFDM.

- Using the same antenna, how much more RF power would a COFDM transmitter need to generate to replicate the same coverage as an 8VSB transmitter?

Ignoring interference limitations, the transmitter size would have to be increased by four times (or +6dB) to maintain the same coverage in the fringe of the service area. The receiver noise threshold is based on average power and COFDM requires 4dB more average power near the receiver threshold than 8VSB. Changing out only the 8VSB exciter will result in 2dB less transmitter power output plus the 4dB threshold penalty, for a total receiver threshold loss of 6dB.

- What about interference protection ratios?

COFDM requires greater digital-to-digital and digital-to-analog interference protection ratios than 8VSB. The SET/ABERT field test data supports the Harris analysis that co-channel COFDM-to-COFDM will require about 4.8dB more protection than 8VSB-to-8VSB. Co-channel COFDM into analog interference will require about 1.0dB more protection than 8VSB into analog. Considering the tighter co-channel interference protection ratios required by COFDM, if the transmitter power was increased to provide the same coverage as 8VSB, the FCC table of channel allotments would have to change to maintain the current protection ratios. Even if the transmitter power was not increased, acceptable protection ratios cannot be met with COFDM. This will create large areas of interference where the signal cannot be received and could significantly delay the introduction of DTV within the U.S.

- Would changing to COFDM increase AC power costs?

Yes. The AC power costs for a 50kW 8VSB IOT transmitter operating 24hrs/day would typically be about $175K/yr. To replicate coverage, a 200kW COFDM IOT transmitter would be needed, with typical power costs of $586K/yr. The European approach to broadcasting, which uses many low power transmitters, is not impacted as much by the higher peak-to-average ratio required by COFDM.

- Are there other long-term issues to consider if a station needs to make the transmitter purchase decision now?

Yes. If the broadcaster wants to replicate analog coverage as the FCC planned for 8VSB, the digital transmission plant would have to provide about 6dB more peak effective radiated power (ERP) for COFDM. This could be accomplished by planning for a larger transmitter, a higher gain antenna, or a combination of both.

If the FCC were to make the decision to change to COFDM from 8-VSB, the final impact might make that perfect storm of recent fame look like a gentle summer zephyr.

In the simplest of scenarios, existing DTV stations would only have to replace their encoders and exciters. Not a trivial amount of cash, but nothing that would put them off the air. The only problem is that they would sacrifice a great deal of coverage if there were no additional change to the facility. First, look at the ratio of peak to average power for the two systems. For 8VSB, the ratio is 4.5 to 6.5. With current encoding/compression practices, the number of four has been found to be livable. For COFDM, we have been advised that the measured value is closer to 15. Even if this were to be held down to eight by improved encoding or compression methods, the transmitter would have to be doubled in peak output power capacity to maintain the same average power.

Next, the carrier-to-noise ratio is different for satisfactory COFDM reception. This calls for another 4dB in signal level to maintain the same service area. The overall result is a necessary increase by a factor of at least five to maintain the same service area. If the station is now running a transmitter with 100KW peak power output, the increase would be to 500KW peak power output. The power bill would be interesting, but that big long row of IOTs would be pretty.

On the good side, based on some of the highly publicized tests, there would be a significant improvement in reception in moving vehicles and when using simple non-directional antenna systems. If you have an overpowering need to download a stock market quote to your palm pilot, you have a winner. On the other hand, without a significant increase in the transmitter facilities, kiss your rural audience good-by. Unless that increase in transmitter power is made, the station's service area will decrease significantly. The smaller audience would no doubt show up in what advertisers would pay, especially those in the agricultural industries who have to reach the rural markets. However, those wanting to deliver data to mobile users may pay more. Just think, by watching a dashboard display, it would be possible for a driver to crash at the same time as he sees his company's stock do the same.

Next, what about the allocation scheme for the country? The good news is that one manufacturer advises that COFDM does not cause a serious change in interference when the power levels are the same. Interference appears as an increase in background noise rather than a more annoying beat pattern. However, when the needed power increase is taken into account, the amount of interference received by NTSC stations will increase by several dB. If that isn't acceptable, the allocation plan will have to be revisited and new channels assigned. For a station that has already purchased a new antenna and where that antenna is a slot type radiator, it will probably be necessary to replace the antenna. The transmission line will be satisfactory unless the power increase exceeds the rating of the line, in which case the used equipment vendors will be extremely happy. Forget about selling the old antenna. Unless you happen to find someone who wants the same pattern for the same frequency, a used slot antenna becomes an interesting lawn ornament to place behind the transmitter building next to the model of the color wheel TV receiver the old chief engineer made years ago.

If the station is willing to accept the reduced service area, they will have a whiz-bang signal into portable receivers and cars - sort of. One report indicates that COFDM does a wonderful job of compensating for reflections with a delay up to about 15s. Above that, it crashes. As an example, stations transmitting from the Hancock building in Chicago have a reflection of more than 15s in their received signal to the area northwest of the Sears Building. Does that mean that there will be a dark area in a nice sized wedge north of O'Hare Field? It probably wouldn't be anything to worry about as it shouldn't involve more than 500,000 viewers. There still would be plenty left, especially when all the cars are added up. If this is problem is found to be a reality, it should be addressed prior to changing the system. It's possible that it can be solved by the coding geeks in the back room. That still leaves the significant problem of the required power increase to be solved.

On the other hand, this writer is not a policy maker but only a simple country engineer. Let the big boys figure out what to do and we will live with it. But please quit second-guessing every decision.